Text Sentiment Analysis Based On Bert-BiGRU-CNN

Emotion is a subjective consciousness that computers do not have,it is an important step in the field of artificial intelligence.Text Sentiment analysis is to judge the emotional tendency of critical text with emotional color,it needs to consider emotional feelings rather than simple logical calculation.For the field of artificial intelligence,having the ability of Sentiment analysis is a qualitative leap.With the development of the Internet,the potential value of online corpus has been widely concerned by researchers from all walks of life,which promotes the development of the research field of emotional analysis of critical texts.At present,there are still deficiencies in text Sentiment analysis methods.By analyzing the research background and current situation of Sentiment analysis,this paper proposes a text Sentiment analysis method based on Bert-Bi GRU-CNN,and designs and implements a visualization system based on this model.The main research contents include the following two aspects:(1)Text Sentiment analysis method based on Bert-Bi GRU-CNNThis paper adopts the method based on deep learning,takes the vector representation and feature extraction of text as the research point,and proposes a text Sentiment analysis method based on Bert-Bi GRU-CNN.Bert-Bi GRU-CNN text emotion model uses Bert to get the dynamic word vector,which solves the influence of polysemy on text Sentiment analysis;The global features of long text sequences are obtained by Bi GRU,and the local key features are extracted by CNN to improve the feature extraction performance of the model.Experimental tests on hotel reviews and online shopping data sets show that the combination of Bert,Bi GRU and CNN for text Sentiment analysis has a certain improvement in classification performance compared with the baseline model,indicating the effectiveness of the text Sentiment analysis method based on Bert-Bi GRU-CNN.(2)Design and implementation of text Sentiment analysis visualization system based on Bert-Bi GRU-CNNIn order to show the results of text Sentiment analysis more intuitively and clearly,combined with practical application,a text Sentiment analysis visualization system based on Bert-Bi GRU-CNN is designed and implemented.From the perspective of system function and visualization,the idea and structure of the system are designed in detail to ensure the integrity and availability of the functions of each part of the system.The background of the visualization system is realized on the basis of the lightweight development framework Django.At the same time,the web interface is developed by using HTML,CSS,Java Script and other technologies.The system interface is friendly,easy to operate and has certain practical application value.